Short-term traffic flow and speed estimation with traffic data obtained using the internet of things technology

Intelligent transportation systems (ITS) are defined as transportation systems equipped with smart sensing, computation and communication technologies to increase operational efficiency and capacity. With the technologies utilized within ITS, accurate, reliable and real-time data about the traffic components are obtained. Thus, real-time traffic management, control strategies, delay, congestion and energy consumption are reduced in transportation systems. Therefore, accurate and real-time traffic estimates are a critical need for ITS to work effectively. In this thesis, it is aimed to make short-term traffic speed and flow estimation using real-time traffic data obtained thanks to the internet of things (IoT) technology used in ITS applications. Accordingly, ITS firstly has been described in detail and its architecture and historical development have been explained. Usage areas and examples in our country and in the world are examined. Afterward, IoT technology is defined and its architectural structure is mentioned. IoT technology applications in ITS applications are explained with examples. Finally, within the scope of the thesis, short-term traffic speed and flow are estimated by using the data obtained from the sensors in the Performance Measurement System (PeMS) of California Transportation Department (Caltrans). Among machine learning algorithms, artificial neural networks and support vector machine methods are used to make these estimations. Traffic speed and flow values for 15-minutes ahead are estimated as the short-term values. In addition, scatter plots showing the speed-flow value relationship are created using both real data and estimated values. On the basis of these scatter plots, the congestion situation of the traffic are assessed based on the three-regime speed-flow model.